本研究提出一套藉由Kinect遠端控制機械手臂與自走車,目的是可以應用在危險的環境中執行任務,此自走車上另有安裝鏡頭,讓使用者可利用攝影鏡頭即時遠距離操控。本研究方法以Kinect取代普遍的攝影機,Kinect具有深度鏡頭可獲得深度資訊,利用影像辨識可以區分出人與物,再利用Kinect的彩色鏡頭去偵測人的膚色辨識出人的肢體動作。得到肢體動作資訊後,運用TCP/IP架構下傳送控制數據至機械手臂與自走車上,以達到即時控制之目的,使得機械手臂能即時做出操作者的動作並且能準確夾取目標物;另一方面操作者也可使用鍵盤下達指令,將指令透過RS232傳送到自走車上的BS2晶片,使得自走車也可以即時控制,並到達準確的目標位置。
This thesis proposes a set of remote mobile robot control using Kinect. The goal of the thesis is to apply to perform tasks in the dangerous circumstances. A camera is installed on the mobile robot, so the operator can timely use the remote-control through the camera. The method of the thesis is using the Kinect to replace the generic camera. The Kinect has a depth-camera, which can capture the depth information and differentiate between human beings and the objects by the image recognition. In addition, the Kinect has a color-camera, which can capture the skin color to recognize the body action of human beings. After the information of body actions is gotten, the control data is sent into the robot arm and the mobile robot by the TCP/IP. Finally, the robot arm can timely imitate the operator’s actions, and it can exactly clamp the target. On the other hand, the operator can also use the keyboard to input the commands and the commands are sent into the BS2 chip on the mobile robot through the RS232. The mobile robot can also timely be controlled that it can exactly move to the target location.